Download Full Text (1.3 MB)

Publication Date



Infant acute lymphoblastic leukemia (ALL) is an especially aggressive form of leukemia, particularly among patients with a rearrangement in the KMT2A gene (KMT2A-r). Approximately two-thirds of infants with KMT2A-r ALL relapse within a year of initial diagnosis, with relapsed disease particularly resistant to treatment and associated with high mortality. The ability to better predict which patients are likely to relapse, especially relative to current residual disease detection methods like flow cytometry, would greatly improve treatment selection and patient outcome. In this study, we set out to identify novel prognostic markers in patient data at diagnosis. We utilized 10X Genomics single-cell RNA sequencing (scRNAseq) of bone marrow or peripheral blood samples with leukemic cells at diagnosis and compared gene expression between infants with KMT2A-r ALL who later went on to relapse (n = 19) versus those who did not (n = 6). Differential expression analysis identified novel cancer-linked genes associated with future relapse (n = 484), including NR5A2, TP53BP2, and BCL2L11, or lack of future relapse (n = 461), including CSMD1 and SCAI. Taking advantage of the single-cell resolution of our dataset, we also applied a recently published tool for assigning scores predicting resistance or sensitivity to treatment for individual cells. Consistent with the literature, patients in our dataset who went on to relapse had a significantly higher proportion of treatment-resistant cells at diagnosis than those who did not go on to relapse, further suggesting the prognostic value of this metric. Overall, this study leveraged scRNAseq data to identify multiple tools that can be used to predict relapse risk in infants with KMT2A-r ALL. These findings should enable better patient risk stratification and therapy selection at diagnosis for infants with this aggressive disease and may also lead to the development of novel individualized treatments in the future.

Document Type


Single-Cell Genomic Data Reveals Heterogeneity In Infant Acute Lymphoblastic Leukemia